Struggling with the Praxis of Social Accounting: Stakeholders, accountability, audits and procedures[1] by:

Rob Gray & Colin Dey (CSEAR, University of Dundee) Dave Owen (University of Sheffield) Richard Evans (Traidcraft Exchange) Simon Zadek (New Economics Foundation)

DRAFT 4D: January 1997 Address for correspondence: Professor R.H.Gray, The Centre for Social and Environmental Accounting Research, Department of Accountancy and Business Finance, University of Dundee, Dundee, DD1 4HN, Scotland.

Abstract This paper is motivated by three related, though not entirely congruent, aims. The first aim is to seek toinitiate moves towards a "normative theory" - a conceptual framework - for the developing of social accounting by organisations. The second is an inductive aim - namely to attempt to draw out best practice from a range of social accounting experiments - illustrated, in particular, by reference to two short cases from Traidcraft plc and Traidcraft Exchange. The third aim is to draw from the conclusions reached in the exploration of the first two aims and attempt to identify any clear "social accounting standards" or "generally acceptable social accounting principles" which can be used to guide the new and emerging social accounting practice. In addition, the paper has a number of subtexts. The more obvious of these relate to attempts to link back to the more critical accounting literature's critiques of social accounting; to address the tension between academic theorising and engaging with practice; to synthesise different approaches to social accounting practice; and to respond to the urgency that the recent upsurge in interest in social accounting places on the newly formed Institute of Social and Ethical Accountability. The paper is, therefore, an ambitious one. The breadth of this ambition ensures that coverage of issues must be thinner than might typically expected in a paper of this sort. But it is explicitly exploratory and, rather than provide answers, it seeks to offer a collective view from experience and encourage engagement with the rapidly evolving social accounting agenda.

1. Introduction Social accounting, after its brief heyday in the early to mid-1970s (see, for example, Estes, 1976; AICPA, 1977; Gray et al. 1987; 1996a), had disappeared almost entirely from both the language of practice and the orthodoxy of conventional accounting by the late 1970s. It returned to its previous obscurity even in accounting academe, (see, for example, Gray et al., 1987; 1991; Parker, 1986). The re-emergence of environmental accounting in the late 1980s and early 1990s, (in so far as environmental accounting can perhaps be conceived of as one part of the social accounting concern[2]), appears to have been at least one of the reasons behind a current resurgence of interest in social accounting[3]. By the mid 1990s, not only was there an increase in academic attention given to the area but, more importantly for this paper, there was a notable re-emergence of practice in the field. Indeed, social accounting is attracting an almost unprecedented level of interest at the present time. Recent social accounts issued by Traidcraft plc, Shared Earth, The New Economics Foundation and The Body Shop in the UK, the SbN Bank in Denmark and Ben and Jerry's in the USA, for example, have all attracted widespread publicity. The Royal Society of Arts in the UK has published An Inquiry into Tomorrow's Company which recommends the development, use and disclosure of social performance indicators. These and related developments[4], together with the recent formation of the Institute of Social and Ethical AccountAbility (ISEA)[5], have made the technical problems of social accounting a matter of some urgency if social accounting is to develop in any systematic way and neither fizzle out through lack of direction or be captured and trivialised by powerful organisations. This emergence of practice has brought both an increase in the examples from which researchers could learn about social accounting practice as well as an urgency to the question "how does one do social accounting?". How to do social accounting should not be a trivial question. If organisations are permitted to develop methods of social accounting without any critical assessment of their activities then any emerging "best practice" can be expected to be partial, ad hoc, immanent and legitimising - thereby falling foul of the marxian, deep green and feminist critiques of social accounting. (see, for example, Tinker et.al., 1991; Puxty, 1991; Cooper, 1992; Maunders and Burritt, 1991). Therefore, what might a "good" or even "ideal type" of social accounting look like? Once we have an idea what it might look like, how might this be developed as an applied practice? Such questions, in turn raise two fundamental problems: what is social accounting? and what is the theoretical, political and ethical framework within which ones' answers are to be framed? Not only is there the tautological concern that assessing "good" social accounting needs a yardstick against which it might be judged but the justification for deriving a practice must lie in the values and emancipatory moment which underpin the suggestions for practice. Social accounting academics have long found themselves in this difficult and relatively under-populated area - lying between sophisticated critiques of current practice, imagining new (and "ideal") accounting systems and actively engaging with (hopefully emancipatory) practice[6]. The need to resolve the tensions that result from this separation of the "academic" and the "practical" is an increasingly important theme in the accounting literature (see, for example, Mitchell et al, 1991; Sikka, 1987; Sikka et al, 1989; 1991; Tinker et al, 1991; Gray et al. 1988; 1991; 1996a; Gray, 1992)[7]. It is for this reason that we use the term praxis in the title of this paper. Whilst the theoretical critiques of accounting and new accountings must, in the interests of scholarship, be engaged with fully, this is not enough. Practice must be encouraged and we must find ways to encourage that practice to develop in a manner which is potentially

1 of 18 emancipatory - not repressive[8]. It is these tensions that this paper seeks to address. The paper is, therefore, very - perhaps overly - ambitious. To theorise a social accounting, to deduce lessons of best practice and to begin to derive standards for social accounting practice - whilst remaining cognisant of the radical critiques of social accounting is clearly more than a single paper could hope to achieve. At the same time, the practice of social accounting is developing rapidly as we write. New methods of accounting are being derived, new institutional arrangements are being developed and practitioners are meeting new difficulties. The paper is therefore explicitly exploratory and excuses its thinness in places in its attempts to bring some system to developing practice as a matter of some urgency. If the paper can then act as a catalyst, research agenda and starting point for further refinement, it will have been more than successful. This paper is part of continuing series in which we seek to develop our understanding of the theoretical and practical issues of social accounting (see, for example, Evans, 1991; Zadek and Evans, 1993; Raynard, 1995; Zadek and Raynard, 1995; Dey et.al., 1995; Harte and Owen, 1987; Gray et al., 1988; Gray et al, 1996a). In line with those other papers, the present discussion is based on a belief in the primacy of democracy[9]; the importance and power of accountability in the development and discharge of democracy; and is empirically grounded in the experience arising from the production of the social accounts mentioned above[10]. The paper itself seeks a synthesis of the different approaches to social accounting and then explores how the resultant theoretical structure can be employed in the systematic production of "auditable" social accounts[11]. This exploration is undertaken through illustration from two cases based on the UK Traidcraft experience. This process, we believe, takes us towards a "portable" social accounting methodology that allows us (to use conventional accounting terminology) to begin a process of iteration towards a conceptual framework for social accounting and the derivation of Generally Accepted Social Accounting Principles and Basic Social Accounting Standards[12]. The paper is structured as follows. Section 2 introduces some of the underlying themes that this paper draws from. Rather than an analysis of these themes, this section simply attempts to locate the discussion in the wider theoretical literature. Section 3 then looks at terminology and outlines some different approaches to social accounting. Sections 4 and 5 examine three of the theoretical perspectives which can be taken on the organisation-society relationships and which, thus, can employed to guide the form of any subsequent social account. A broad synthesis of these three perspectives is offered in Section 6. Sections 7 and 8 report the experiences of, respectively, Traidcraft plc and Traidcraft Exchange in their development of their own social accounts. Lessons about constructing the (social) accounting entity are drawn from the first case. The second case is used to illustrate how "three layers of information" can bring us towards a more complete social account. This case also flags up the role of auditing. Section 9 is a short report of the key elements that our experiences have led us to believe are the backbone of a systematic social accounting practice and might, thus, form a basis for the emergence of social accounting standards.

2. Locating the Discussion: Some Initial Theoretical Reflections It would be inappropriate - if only for reasons of space - to attempt to revisit all aspects of the social accounting debate here, (but see Gray et al.,1996a for an introduction to some of the issues). This section will, therefore, seek to identify - in very general terms - the principal theoretical themes which will underpin the later analysis and discussion. In the first place, social accounting is conceived of as the universe of all possible accountings. From a conventional accounting perspective, it is what happens if the constraining principles of conventional accounting (i.e. an accounting entity, a focus on economic events, financial description of those events and an assumption of (predominantly financial) users - see Gray et al, 1996b) are relaxed. This universe is, probably, infinite and, consequently, beyond our scope here. So we initially restrict our discussion to formal accounts from accounting entities. (These issues are discussed in more detail below). We thus exclude - most obviously - informal (verbal and non-verbal accounts) between individuals and groups. We similarly restrict ourselves to a concern with the production of single periodic accounts. We should, however, stress that such single accounts do not assume a single point of view. That is, the social accounts are assumed to be, at least potentially, both polyvocal and multiple (see, for example, Zadek and Evans, 1993). This restriction to a single, formal account may well open the resultant social accounting to criticisms from feminist - and perhaps post-modern - critiques. That is, our approach to the social account may be masculine/gendered and modern. We have not addressed this formally, but the later discussion does, we suggest, acknowledge these charges and may go some way towards meeting the criticisms. Furthermore, there are both political and social issues embedded in the social accounting which we have not explicitly explored. For example, almost any system of social accounting of which one can conceive will involve transfers of power either through the democratisation of organisations and of capital or through the capture and manipulation of social issues by organisations and by capital. Similarly, there are a wide range of concerns over such matters as the role of accounts in reifying, constructing, reinforcing and trivialising organisations and relationships both within and between those organisations, (see, for example, Hines, 1988; 1989; 1991; 1992; Francis, 1990; Lavoie, 1987; Meyer, 1986). How successfully such critiques can be levelled at this project - and perhaps undermine it - are matters dealt with only tangentially in this paper. Next, social accounting is conceived here as being both hermeneutic and emancipatory. It is hermeneutic in the sense of a "hermeneutic dialectic process" (Laughlin and Broadbent, 1996) which seeks out a means that might lead us towards reflexive mutual understanding between the organisation and its stakeholders. These understandings will relate to both organisational activity and the accountability(ies) - see later - attaching to those activities. In this sense, the social accounting practice must be continuing and evolving. It is emancipatory in that its aim is to redress power asymmetries between organisations and their stakeholders through the reporting of information. This emancipatory moment lies beyond the accountability itself and, rather, underpins the project and provides its motivation. The project seeks to enhance the democratic virtues of transparency and accountability. It seeks this through, (admittedly under-specified), assumptions that an informed demos is thereby empowered in its decision-making and action to seek more benign organisational activity. Similarly, the production of social accounts is implicitly assumed to have an information inductance effect on the part of organisational managers that will encourage more ethically desirable forms of activity. However, the specific focus employed here is rather less dependent upon the optimism (and utilitarianism) which would justify a social accounting by reference to outcomes flowing from that accounting information but is rather more concerned with deontological and intrinsic motivations. That is, the fulcrum of social accounting employed here is the discharge of organisational accountability. Accountability is conceived of as relating to the rights to information of a participatory democratic society which, for the sake of this

2 of 18 project, is conceived of as neo-pluralist in structure[13]. Such a series of choices have two implications. First, the choice of a neo-pluralist democratic accountability defines a "problem space" within which all parties can debate. It is reminiscent of - though in no sense synonymous with - Habermas' ideal speech situation. That is, western practice and western critique can all engage in this space because the conception does not necessarily deny the legitimacy of organisational practice but brings to it critiques drawn from pristine liberal economic democracy, marxism, deep green and feminist positions. Whilst social accounting exists in a space critiqued by each of these positions - whether the critique from practice or the radical critique from academe - there remains a hermeneutic engagement, which a priori, we find valuable - even essential[14]. The second consequence of our choice is that, in a strict interpretation of accountability, action and decisions as a consequence of the discharged accountability are not necessarily part of the story. A participatory demos may well have rights to information but once informed may choose not to act on that information. The assumptions of action and decision lie prior to the concern with accountability itself. With this outline of the prior concerns which attempt to link this project with the wider theoretical literature, we can now turn to the body of the paper and explore the issue of terminology and definition before moving into an exploration of the conceptual frameworks for an applied social accounting.

3. Social? Ethical? Accounting? Auditing? A very wide variety of terms have been employed in social accounting. Historically the terms "social audit", "social responsibility accounting", "corporate social reporting" have, from time to time, been popular (see, for example, Mathews, 1993) whilst more recently social accounts have appeared under the titles, inter alia, "social audits" (the early Traidcraft reports for example), "ethical statements" (SbN Bank) and "values report" and "social statement" (Body Shop). Further analogues can be found - "ethical audits" is a term favoured by some individuals, for example[15]. The confusion generated by the diversity of terms is partly illusory but it also partly real. All the discussion and practice holds a number of central ideas in common but, through the use of a diverse terminology, holds these central ideas with different degrees of exactitude. Indeed, the different conceptions permit undefined terms to encourage agreement and also, perhaps, to avoid historical baggage attached to terms which parties may prefer to avoid[16]. Nevertheless, we stay with social accounting as the generic term because it is the longest established and the simplest term (construct) with which to work. Furthermore, as we shall see, we find its resonances of conventional accounting especially useful when defining the elements of a social accounting process[17].

All social accounting (regardless of its actual title) relates to the presentation of information about organisational[18] activity. The resultant account (in whatever form that might take) is presented to someone - even if it is only the organisation itself, (see below). The social account, as with any other account, presupposes some defined accounting entity[19] (there has to be something about which to account). Equally, there is an implication of complexity and size in the relationship between the accounting entity and those to whom the account is presented. This last point arises because we tend to find that formal accounts are only necessary in complex situations in which personal communication, trust and intimacy (what Rawls calls "closeness") is threatened or has disappeared. (For more detail see, for example, Lehman, 1995). In personal, usually relatively simple situations involving only a small number of parties and/or activities, an informal account will suffice. The precise point at which an informal accounting in or by (for example) families breaks down and leads to the need for a formal account in or by (for example) multinational companies is quite unclear but deserving of further research. Finally, there is, implicit in the notion of an account, some intended or actual recipient to whom the account is (informally, perhaps) addressed. Within the neo-pluralist accountability framework employed here, the stakeholders are those with rights to the account and it is for them that the account is prepared. Whether or not they use it, and if so for what, and whether or not other parties see and/or use the account are largely irrelevant[20]. So, beyond the identification, prioritisation and needs of stakeholders (see below), it is possible to leave any conception of the recipient's decision-making or choices about action as implicit in the discussion. These three elements - the accounting entity, the account and the intended recipient (or group of recipients) of the account - are not sufficient in themselves to define the social accounting. Figure 1 provides a collection of possible accounting processes which involve these three elements. The possible groupings of social accounts are initially defined by reference to who prepares and who receives the account.

FIGURE 1: A Categorisation of Social and Environmental Accounting and ' Audits'

Report for the consumption of... Report compiled by....

Internal participants External participants

Internal participants 1. 2.

Environmental audits/accounting; Regulators' report - e.g. EPA; EMS - EMAS/ISO14001; Supplier audits; Attitude audits; Duty of Care audits; Stakeholder testing; Environmental consultants; Compliance audit - e.g SA8000; Social responsibility checks; Social responsibility audit; Market/stakeholder research Mission/values audit; Image audit

External participants 4. 3.

Disclosure in annual reports; The ' external social audits'; The ' silent social accounts'; Ethical investment/EIRIS; Environmental reports; Consumer audits; Social reports; Pressure group audits; GRI/' Sustainability' reports; Environmental/Greenpeace

3 of 18 Compliance reports; Social Audit Ltd; Mission/values statement; Journalists; Adverts/stakeholder education; Competitors; NGO social audits Trades union reports;

Adapted from R.H.Gray (1991) Trends in Corporate Social and Environmental Accounting (London: British Institute of Management) We are going to restrict ourselves to the formal account and, furthermore, to the formal account which is prepared by (or on behalf of - see below) the accounting entity and (externally) reported or disclosed to parties other than the directors/controllers of the reporting organisation, (although this does not preclude such parties from receiving - and, indeed, using - such an account). (This restriction places us in the shaded box in Figure 1). The reasons for these restrictions are not, we believe, trivial. The restriction to the formal account arises from the complexity of the organisations about which the questions of social accounting most obviously arise[21]. The restriction to externally disclosed reports is driven by the commitment to accountability - of which more later, (see, for example, Medawar, 1976). The matter of who prepares the report is really a matter of pragmatic[22] and economic concern. Whilst, for example, the first Traidcraft report (1993) was prepared with considerable external input from NEF, it was ultimately published by - and thus ' owned by' - the reporting entity. This ownership is important because the organisation must thus stand by the contents of the report - it is, after all, the duty of the organisation to be accountable that has driven the account in the first place. But there are other pragmatic reasons - most especially that (i) externally prepared reports typically experience problems in gaining access to the appropriate levels of data and (ii) the economics of the regular production of social accounts will effectively preclude any systematic mechanism other than self-reporting by the organisation, (see, for example, Medawar, 1976; Gray et al., 1987; 1996a). Furthermore, the analogue of relating social accounting to conventional financial accounting[23] naturally leads us to assume the preparation of the report by the accounting entity. These restrictions, whilst they place us in the shaded box in Figure 1, still do not, of themselves, produce a full definition of the social account. As Figure 1 might suggest, a wide range of fairly basic examples - as well as the more developed examples we have already referred to - would still be encompassed by our terms. Indeed, most large companies produce a range of social and environmental information which, if collated, might pass as a social account, (see, for example, Gray, forthcoming). Consequentially, it might be helpful if a further schema, refining slightly the elements of Figure FIGURE 2: A SCHEMA OF SOCIAL ACCOUNTING 1, was presented at this point. Figure 2 provides a classification of approaches to social APPROACHES accounting based on the complexity and system of the approaches and, more pertinently, 1. Company Preferred: their likely desirability to a conventional business organisation[24]. The terms will be discussed existing corporate and developed as the paper progresses. It is the second group - the systematic social reporting of both accounts - that is of concern to us in this paper. To further refine our definition we need to voluntary and mandatory address the matter of theory and, in particular, the place and interpretation of accountability in data; the derivation of a social account. existing corporate public relations and 4. Accountability and Stakeholder Perspectives advertising/education Our experience with social accounting plus our reading of the literature lead us to conclude data; that there are three dominant ways of theorising[25] the (accountability) relationship between collation of the above into an accounting entity and its "outside world"[26]. We want to conceive of these theoretical a single "social perspectives as a series of (overlapping) layers of perspective which can be synthesised and accounting" document - built up into a rich conception of the organisation-society interaction. These three the "silent account"; perspectives are a stakeholder perspective, an accountability perspective, and, what we shall "one-off" experiments with have to call a polyvocal citizenship perspective (PCP)[27]. These three conceptions build up approaches to social from the "harder", more functional, organisation-centred stakeholder perspective, through the accounting; slightly "softer", society-centred accountability perspective to (perhaps) the "softest", 2. Systematic Corporate Social stakeholder-centred polyvocal citizenship perspective. (We will discuss the first two below Accounts: and then examine PCP in more detail in the next section). stakeholder reporting; Stakeholder perspective: The stakeholder approach to analysis is well-established in the describing the management (and accounting) literature, (see, for example, Ullmann, 1985; Donaldson, 1988; characteristics of the Roberts, 1992; Nasi, 1995). Its essence is the definition of all those groups or parties[28] who stakeholder are influenced by and/or who influence the organisation (or accounting entity). From this point relationship(s); on, stakeholder theory struggles to maintain anything other than an organisation-centred accountability reporting; legitimacy because whilst the groups may be defined with a fair degree of objectivity, who reporting of the voices of (other than the organisation) is left to define the priorities amongst the stakeholders and the the stakeholders; information that should be disclosed to each one? Stakeholder theory, therefore, is typically concerned with how the organisation manages its stakeholders. Thus, information disclosed 3. External "Social Audits": to the stakeholders may more properly be assumed to be part of a legitimacy and/or social single issue reporting construction process by the organisation. Stakeholder theory is relatively silent on how the (e.g. consumers, organisation does - if at all - monitor and respond to the needs of the stakeholders. It will do environmentalists etc.); so, generally speaking, when it is in the organisation's traditional interests (profit-seeking for example) to do so. Therefore a social account based upon the stakeholder perspective only systematic social audits; has social value if we assume the beneficence of the organisation and further assume that one-off reporting the stakeholders' needs can be morally subsumed within those of the organisation[29]. If we (e.g.investigative assume this, then "market forces" will generally produce the sort of social accounting which is journalism). in the organisation's best interests. We might reasonably assume that it is this thinking which produces the sort of voluntary social and environmental disclosure we currently see. Despite its serious limitations, stakeholder theory does help us. It defines the influencing/influenced groups for us and explicitly defines what accountability the organisation itself is willing to recognise and discharge[30]. To deny the organisation any role in a definition of a

4 of 18 social account seems inappropriate and largely indefensible. This, therefore provides our first layer - the stakeholder analysis within which the organisation defines the accountability. To this we now add the accountability perspective - society's views. Accountability perspective: Accountability and its role in social accounting has been widely discussed. It is concerned with the relationships between groups, individuals, organisations and the rights to information that such relationships entail. Simply stated, accountability is the duty to provide an account of the actions for which one is held responsible. (For more detail see, for example, Gray et al., 1986; 1987; 1988; 1991; 1996a). The nature of the relationships - and the attendant rights to information - are contextually determined by the society in which the relationship occurs. It is definitionally true that some sort of relationship will exist between an organisation and each of its stakeholders. As we saw above with stakeholder theory, part of these relationships may be economic in nature and the terms determined by the parties - as reflecting their relative power in the relationships. The information flowing through the relationship will be determined by the power of the parties to demand it (a power which, where it exists, could arise from either the intrinsic abilities and power of the groups concerned or from the legislative processes of the society) and/or the willingness/desire of the organisation to provide it. Society-as-a-whole reflects (what might be thought of as) a concern that all such relationships and their attendant information rights should not be left entirely to the parties. The most obvious manifestation of this is statute law (e.g. companies acts, equal opportunities legislation) and standards established by statutory bodies (e.g an environmental protection agency, health and safety at work inspectorate). Additionally, other mechanisms (such as, for example, voluntary codes of practice) will from time enter the public domain as an agreed or, at least, negotiated part of the stakeholder relationship. These ' empirical, beyond-law' determinants of accountability have been referred as ' quasi-law', (see, for example, Gray et al., 1987; 1988; 1991; 1996a). The extant law plus this ' quasi-law' will therefore represent the first (and major) element in the construction of society's views on the accountability of organisations, (see, for example, Stone, 1975). It is, of course, naïve to assume a simple one-to-one mapping of a "society's" beliefs about the nature of relationships and attendant information rights and extant law - even with addition of ' quasi-law'. On the one hand, rights to information must reflect assymetries of power and essential lags between a society's views and the enactment of law (see, especially, Stone, 1975; Epstein and Votaw, 1978; Dowling and Pfeffer, 1975). On the other hand, such rights to information can be argued to comprise both ' positive' (legal) and ' normative' (moral) rights (see, for example, Likierman, 1986; Likierman and Creasey, 1985). Empirical determination of the moral rights is clearly contestable. These moral rights must, in some manner, be added to the positive rights to reflect current views of accountability. Some of these rights will be accepted by organisations - through, for example, mission and/or ethical statements - and, to this extent at least, there will be a convenient overlap between the information rights determined by stakeholder theory and those determined by accountability theory[31]. But many moral rights will not be so accepted. Here one must resort, within an accountability framework, to other expressions of society's values. Malachowski (1990), for example, argues for a recognition of media concern as an approximation of new emerging issues of society's concern. Broadly, this may be a plausible basis. For illustration, health and safety/concern over accidents rose as a major issue of concern in the UK in the 1980 whilst environmental issues clearly claimed a place in society's conception of the organisation-society relationship in the early 1990s, (see, for example, Gray, Kouhy and Lavers, 1995a). Less obvious issues which are not governed by extant law, - for example, fair trade and impact on developing countries - are more clearly a matter for negotiation. The second strand of social disclosure would therefore derive from the established accountability relationships and seek to provide information to which the stakeholders have rights[32]. There is a danger, however, that a social accounting founded on a combination of stakeholder and accountability perspectives would be too inert and only slowly responsive to changing stakeholder needs. More particularly, the modern basis of these perspectives leave them open to challenges from perspectives arising from critique informed by more fluid and, perhaps, more "post-modern" conceptions. It is concerns such as these that encouraged many of the newer social accounts to adopt the polyvocal citizenship perspective. It is to this that we now turn.

5. New Conceptions of the Organisation-Society Dialogue: A Challenge Many of the more recent social accounts, especially those influenced by NEF and Traidcraft, have adopted an approach to social accounting which has not previously been discussed in the social mainstream accounting literature, (but see, for example, Zadek and Evans, 1993; Raynard, 1995; Zadek and Raynard, 1995). This polyvocal citizenship perspective (PCP) draws broadly from Habermasian discourse ethics and directly from Guba and Lincoln's (1989) Fourth Generation Evaluation[33] and then applies the ideas in a social accounting setting. In its novelty, its explicitly hermeneutic concern and its emphasis on privileging the voices of stakeholders, the PCP offers an alternative conception of social accounting[34]. The approach is built around stakeholder dialogue and its essence lies in providing each of the stakeholders with a "voice" in the organisation. Focus groups are held with each stakeholder group, from which key issues are identified and a wider constituency of the stakeholder group is consulted to collate their views on these and other issues. The social account comprises, predominantly, (but not exclusively), a reporting of the voices of the stakeholders. PCP, thereby, constitutes a different way of seeing the organisation. Thus, if we were to take this to extremes (and thus attempt to typify the PCP approach as a postmodern "straw man"), the terms of the organisation-society/group relationship are established, not by the organisation or the society but by the stakeholders themselves. More precisely, though, PCP is a form of symbolic interactionism (Guba and Lincoln, 1989) in which, (to borrow from the words of Innes et. al., 1996, p6), "the organisational stakeholders are seen as constituting and sustaining their own reality - and that of the organisation - both socially and symbolically. " The PCP social account aims to give voice to these actors so that their systems of interpretation and meanings and processes of structuring and organising are revealed[35].

The PCP, in drawing from Guba and Lincoln, is clearly working at the boundary of the modern and the post-modern[36]. The most important element of this is the way in which the organisation is conceived. Again, taking this to extremes, the PCP might be typified as assuming that the organisation has no existence prior to or independent of the stakeholders. That is, the stakeholders create the organisation in a social constructivist sense. Such a view is central to Guba and Lincoln's analysis which ontologically denies the existence of an objective reality and epistemologically assumes that understanding (in this case represented by the social account) emerges from the interaction between observer and observed, (Guba and Lincoln, 1989, p44). Therefore, whilst the stakeholders are privileged in the construction of the organisation and the social account, the approach is hermeneutic and iterative with refined and developed understandings emerging over several cycles of the social accounting process[37]. The NEF/Traidcraft approach does not stop here though. First, at least in the early attempts at (what were referred to as) "social audits",

5 of 18 the "social auditor" was cast in the role of Guba and Lincoln's evaluator whose job it was to encourage and engage with the dialogue process towards a responsive constructivist evaluation, (in our case a social account rather than an "evaluation" as such). Thus, there is no role for an external, objective audit or attestation because, quite simply, such "externality" and "objectivity" cannot exist within the model. Second, the NEF/Traidcraft approach placed "core values" of the organisation at the heart of the dialogues, (and thus at the heart of the account). These core values - which may emerge from the social accounting process or may be enshrined in mission statements or statements of ethics or principles - become that thing which the social accounting process is "evaluating" and negotiating and - hopefully - changing[38] as a consequence of the stakeholder dialogues. Consequently, the resultant social account contains data reflective of the "core values" of the organisation and which, inevitably, must overlap with the information we would expect to flow into a social account from the stakeholder and accountability perspectives itemised above. Hopefully it is apparent from the brief outline above that the NEF/Traidcraft version of PCP, (what is more usually called the stakeholder dialogue approach), is concerned with praxis. In its concerns to evolve an active practice, it has clearly moved away from the solipsism (see, for example, Laughlin and Broadbent, 1996) of the "pure" Guba and Lincoln approach and, similarly, edged away from the postmodern conceptions of organisational interactions towards something which begins to look more like a Habermas "ideal speech situation", (Habermas, 1978). Thus, within the context of social accounting, it seems that in the PCP "social auditing" process, three things are happening simultaneously: (i) the stakeholders are being encouraged to voice the terms of the accountability relationship - both as they see it currently and how they would wish it to be. That is, the stakeholders are defining the terms of accountability; (ii) the stakeholders are active in defining the accounting entity itself; (iii) the voices of the stakeholders provide an essential element of the basis for the social account of the organisation. Each of these matters needs to be further explored.

6. Synthesising the Perspectives? The first of the above points - that the stakeholders define their own terms of accountability - fits well with the discussion so far. Allowing through the voices of the stakeholders provides a systematic solution to the hitherto unaddressed problem of how to deal with the inevitable limits that an empirical accountability imposes on the social account[39]. So, in terms of this paper, the PCP provides our third (and final) ' layer' of the accountability relationships which comprise the social account:- organisation-, society- and stakeholder- determined terms of accountability. The second point - how one is to conceive of and practically determine the definition of the accounting entity is a subtle and important matter that is all too easily ignored in the stakeholder and accountability perspectives. We will examine this point more fully in a later section. The third of the points - that it is the stakeholders who (in effect) provide the social account - produces both a significant synergy and a significant conflict with other forms of social accounting (as restricted above). The synergy arises from the type of information that the PCP generates. This information includes such matters as employees' levels of satisfaction and motivation, customers' levels of satisfaction, environmentalists' anxieties about the organisation, and so on. This sort of information, whilst it might loosely be thought of as accountability information, is, indeed, both an expression of the stakeholders' voices and management information of the sort that any organisation needs if it is to manage its stakeholders - as suggested by stakeholder theory. In this way, information generated by PCP is both an important addition to the potential discharge of accountability and information which management can use under the stakeholder perspective[40]. On the other hand, the conflict arises between the PCP and more conventional approaches for several reasons. First, under "conventional" forms of social accounting, the organisation provides the social account whereas in PCP the stakeholders provide it, (although the account is actually published and distributed by the organisation itself). This latter point places the PCP (within the more conventional interpretations of social accounting - see, for example, Gray, Owen and Adams, 1996) as a hybrid of self-reporting (the shaded quadrant in Figure 1) and the independent social reporting pioneered by units such as Social Audit Ltd (see the bottom right-hand quadrant of Figure 1). Second, under the extreme version of PCP the account exists only as a social construction of the organisation - the organisation and its relationships do not exists without that construction. The stakeholder/accountability approaches assume an underlying reality which is re-constructed by the social account. That is, the latter suggests that the relationships and rights exist prior to the social accounting. Under PCP the relationships and rights do not so exist and are constructed through the account. Third, whilst the conventional forms of social accounting have been accused of being conservative (see, for example, Tinker et al., 1991), they do have potential for change and evolution (see, for example, Gray et al, 1996; Lehman, 1995; Gallhofer and Haslam, 1995; Power, 1994). They offer this potential by reference to a wider set of societal values that, it is normally assumed, will challenge the organisation's existence. On the face of it, the PCP appears to do the same but may well have more conservative - albeit democratic - tendencies. That is, as a consequence of its genesis in the work of Guba and Lincoln, it may - we emphasise may - be less likely to assume, seek out and expose conflict between organisational legitimacy and stakeholder views. This is, however, a practical, rather than a theoretical problem and relates to the final potential source of conflict with more conventional versions of social accounting. Finally, there is a potential difficulty in that stakeholders, whilst their right to a voice is not, to our mind, contestable, may not be informed in a manner which permits the expression of their voice to challenge the essential problems of organisational legitimacy. Unlike the "external social audits" which raise and extend the boundaries of accountability (see, for example, Gray et al, 1988; Harte and Owen, 1987) the voices may be immanent and only offer comment within the terms already set for them by the organisational hegemony[41]. But these are problems only of degree - all forms of social accounting are susceptible to these problems (see, for example, Puxty, 1986; 1991). Rather, the PCP may, in offering greater synergy with other forms of social accounting, go someway towards muting the criticisms of the social accounting project. Can, then the "conventional" stewardship/accountability approaches to social accounting be synthesised with the PCP approach? The answer is probably that they cannot be fully synthesised - but only because they draw from a fundamentally different ontology. Whilst all

6 of 18 three approaches to social accounting can be thought of as essentially social constructionist, it is the way in which PCP constructs the organisation that raises the essential conflict. Yet it is apparent that both are motivated by similar principal concerns - that is the concerns to develop social accounting and thereby extend accountability and democracy. The two approaches, although ontologically incompatible need not be necessarily mutually exclusive - but rather it seems to us that they represent alternative perspectives on an organisation and are, therefore, mutually reinforcing. There is little question in our minds that the PCP brings an important additional dimension to the processes of social accounting. It adds the, hitherto silent, voices of the stakeholders to the specification and construction of accountability and, as we will see in the following section, helps crystallise the way in which the accounting entity is conceptualised. These, in our judgement, are important developments in the conceptualisation and application of accountability. But for predominantly practical (even pragmatic) reasons, we believe the systematic development of social accounting requires that the organisation be the reporting body. As such, it is the business of the reporting organisation to construct the social account and such a social account, to be complete, must, we infer, recognise the voices of the stakeholders. Under a continuing assumption that the organisation constructs the account then the organisation needs to report (a) that the stakeholders have (or have not) been given a voice and (b) what those voices had to say about the terms of the organisation's accountability. This does not require that the organisation reports the full detail of those voices, (as we see in, for example, The Body Shop Values Report in 1996). Should the organisation choose to report the detail findings from the consultations with stakeholders that then goes beyond the essential requirements of an accountability report. In this way, although we stay grounded in the somewhat more realist ontology of the stakeholder/accountability perspective we open up that perspective in a way which can recognise challenges to that ontology.

Whatever the remaining theoretical conflicts, for reasons of practicability and political expedience, some synthesis needs to be found[42]. The further issues that arise in this reconciliation will become clearer, we suggest, when we examine, first, the construction of the accounting entity and second, when we begin to examine the information which must be contained in the channels of accountability. These follow in subsequent sections. To try and ground the discussion so far, we now turn to two examples of social accounting produced by Traidcraft plc and Traidcraft Exchange. We will discuss the Traidcraft plc case to illustrate the issues arising when we seek to define the accounting entity and the implications this has for the accountability information produced. We will then discuss Traidcraft Exchange as an illustration of how the three layers of social accounting information can be used to work together.

7. The Traidcraft plc Case This section provides a very brief outline of one experience of identifying the organisational stakeholders and defining/negotiating the accounting entity. More detail on the Traidcraft plc experience can be found in Evans (1991); Zadek and Evans (1993), Dey, Evans and Gray (1995) and the published Social Accounts from the company[43]. Traidcraft plc (sometimes referred to as "the plc" hereafter), a small "values-based" UK company working with the New Economics Foundation produced their first comprehensive "Social Audit" in the UK in 1993[44]. This was the first example of a systematic self-generated and reported social account by a UK company and was a remarkable phenomenon for that alone. But the plc went further publishing a booklet explaining their underlying thinking (Zadek and Evans, 1993) and committed the company to regular, developing systematic social accounts of the company in future years. Traidcraft plc has a called-up share capital of £1.8M and an annual turnover of nearly £7M. The company's share capital consists of widely-held - but non-voting - "B" shares, a single "Guardian share" which prevents the takeover or control of the company against the principles of the company, and a body of voting "A" shares which are held by the Traidcraft Foundation Trustees. Traidcraft plc's description of itself is given in Figure 3 which is an extract from the 1994/95 social audit document[45].

The company has about 150 employees and has reported modest profits in the majority of the last few years. Whilst FIGURE 3 ABOUT HERE the company operates in a financial climate and must be economically viable, the relationships it has with its employees and directors, (all of whom accept relatively low salaries), its representatives and agents and its shareholders all ensure that the company is not subject to full blast of the rigours of the late-20th century market economy. The company exists to encourage "fair trade" between the (so-called) developed and developing countries. Most especially, profit and cash-flow are a means to that end not, as in conventional economic and accounting theory, vice-versa. It was not inappropriate therefore that the company should examine the extent to which it was living by its principles - especially as the company has always prided itself on its sensitivity to criticism from its constituents. The 1993 social audit was a joint effort of Traidcraft and NEF with a member of NEF working within the company on the development of the principles of the audit and the collation of the appropriate data, (see Zadek and Evans, 1993). The audit was subject to a degree of independent attestation from an advisory board established by NEF and the report was published in late 1993[46]. The 1993 experience had highlighted two major factors. First, there was confusion of roles over accounting and auditing. Second, it was apparent that producing a systematic and transparent account of the company's activities was more difficult - technically and spiritually - than had been anticipated. The 1994 and 1995 "social audits" were conducted with less involvement from NEF but were still subject to the audit process from the audit advisory board. Traidcraft plc, at the time of writing, have just completed their 1996 social accounts. In the development of the 1995 social account (as it is now called), the plc and the audit advisory board met a real practical difficulty in conceiving of the organisation and its social audit[47]. A systematic heuristic was needed to allow an explicit specification of the elements of the organisation-society-constituents relationship that was practical enough to be applicable to a functioning company and which could be articulated in a way that recognises the societal and political assumptions of the approach in a way sufficiently specified to resonate with the theoretical literature on organisations and accountability[48]. Defining the accounting entity: Prior to the project of which this paper is a part, there was no established method for defining a social accounting system. Equally, there is little likelihood that it is possible to derive any one, unique social account of any organisation[49]. One of the many characteristics that makes the Traidcraft social accounting unique is its ' semiotic nature' - in the sense that the whole process is designed to elicit a shared meaning or expression of the entity known as ' Traidcraft'. Whilst power assymetries are inevitable, all parties who are active in the Traidcraft plc enterprise are given a voice. By this means, a shared account of the organisation begins to emerge. This is not to deny that other interpretations of the account are possible (or even desirable) but rather to suggest that the widely shared expression of the account is essentially an expression of communitarian accountability (see, for example, Pallot, 1987, 1991).

7 of 18 The process is one in which all active parties, including the widely drawn social audit advisory board, the researcher/advisors of The New Economic Foundations (NEF) and the present researchers construct the organisation. The organisation and its accountability: Traidcraft is conceived of as a transparent, permeable membrane organisation where constant negotiation of the organisational boundary takes place (Llewellyn, 1994). It is then conceived of as lying at the centre of a nexus of social relationships which are articulated in a manner akin to a stakeholder model located in a neo-pluralist conception of society. The key here is the notion of relationships. The ' best' social accounts are those which, presumably, best seem to express the essence of those relationships. This conceptual model is illustrated - with a significant degree of simplification - in Figure 4[50].

The relationships are, inevitably, not of equal importance or strength. They are, to a large degree, prioritised by FIGURE 4 ABOUT HERE Traidcraft's own mission statement[51]. The strength - or, indeed, the closeness (Gray ,1992) - of the relationships is also part of the defined organisation. The ' closer relationships' - e.g. with employees, the producer communities and even the representatives/agents - are actually negotiated so that these groups actually become part of ' Traidcraft'. That is, the organisational boundary membrane is especially thin at these points and is managed to accommodate all, or part of, these additional parties. This could be shown in a development of Figure 4 (Figure 5 is a basic attempt at a move towards this) in which, for example, the dominance of the relationship could be shown by the thickness of the line suggesting that relationship and a new ' fuzzy' Traidcraft organisational boundary shown by the wavy line. (Some of this is shown in Figure 5).

The accountability of Traidcraft is only an imposed accountability in the case of any government reporting, any FIGURE 5 ABOUT HERE especial demands from the voting shareholders and the existing financial accountability of all UK companies. All other accountability is assumed because of the moral rights of the accountee parties and/or because of the value-based nature of the organisation and its desire to be accountable, (Zadek and Evans, 1993). This principle is then extended via consultation with the stakeholders to give them an explicit voice - one element of which tells the organisation (and the auditors) whether their initial perception of the relative strengths and "closeness" of the organisational relationships are adequately represented in the conception of the organisational boundary. The organisational boundary - the "accounting entity" - is thus defined by the organisation's, society's and the stakeholders' negotiations of closeness. Those stakeholders which come into the "entity" are then part of the entity and need, consequently, to be included in the account. In this way, what the account is of, and who the account is to is clarified[52]. Additionally, some initial prioritisation of stakeholder relationships is explicitly undertaken. Traidcraft's relationships and the definition of the accounting entity: The 19, broadly prioritised, relationships (shown in Figures 4 and 5) now have to be defined in somewhat greater detail. To provide a full social account, each of the 19 social relationships need to be accounted for. That is, for completeness, the stakeholder, accountability and PCP information needs to be collated, synthesised and reported. However, at the time the Traidcraft social accounts were being prepared this was still an unclear issue. This issue is explicated in the Traidcraft Exchange case below. (Appendix 1 contains an outline of some of the processes that were undertaken in the plc experience with one of its major stakeholders). Discussion of defining the entity: The foregoing has tried to give an indication of the process that was used to begin to define the accounting entity and to prioritise its stakeholders. Hopefully, it is apparent that this was an iterative process driven by a convergence between a conventional stakeholder conception, the accountability perspective and the "softer" PCP conceptions. The articulation of the entity is then used to determine the relationships and the consequent information that the social account should attempt to convey. Undoubtedly, the process was a great deal simpler than it would have been for a more explicitly profit-centred organisation. Traidcraft plc is explicitly "values-centred", is a transparent organisation by nature (with considerable interactions and shared values with many of its stakeholders) and has formally articulated these principles in very extensive mission statements. Furthermore, ideally, the conception of the entity would be explicitly discussed with all stakeholders to ensure that it had semiotic validity. It seems inevitable, however, that such organisational conceptions will be (often implicitly) negotiated and the organisation itself, as the dominant stakeholder, is likely to have a disproportionate influence on the final determination of the entity and, consequently, on the terms of reference offered to the stakeholders in which to express their voices. It is difficult to know, at this stage, how serious a limitation on social accounting this will prove to be.

8. The Traidcraft Exchange Case This section reports a later experiment with the plc's sister organisation - the Traidcraft Exchange - in which a systematic social account was sought employing the three layers of accountability information discussed above. The results of this have been published in the Exchange Social Accounts 1995/1996. Traidcraft Exchange, (occasionally referred to as TX hereafter) is a small charitable company (limited by guarantee) involved in a range of activities primarily concerned with encouraging fair trade. TX has an annual income of a little under £1M (derived through fundraising, grants and contracts) and a staff of between 15 and 20 people - some of whom are "shared" with the plc. It is, thus, a very small organisation. In 1995, TX decided to try and produce its own social account. This would be novel in a number of ways. First, the organisation was so small that whilst data would be easier to capture, it would not be practicable to try and systematise the process through a social bookkeeping system. Second, the definition of the accounting entity would be even more problematic in that (i) the edges of TX and the plc blur at times and whilst this did not seem to be material for the plc, the relative size made it more important for TX; and (ii) voluntary organisations frequently have much less formal organisational interactions and the "membrane" was therefore likely to be thinner, fuzzier and more likely to float back and forth. Third, the experience of the plc had shown that despite the dominance of the PCP approach, the final social accounts had to contain other data which did not emerge from the voices. These, it seemed in discussions, increased the importance of "completeness" that is central to constructing the social account. Finally, it was judged to be a good time to try and discover whether it was now possible to derive a "portable" social accounting process. To this end it was decided that the social accounting process should not rely on NEF, should be undertaken by a member of TX with no previous experience of social accounting (although consultancy advice was available from the authors) and, perhaps most dramatically, the statutory financial auditors would be requested to publish an audit opinion on the social accounts. This final point - the statutory financial auditors expressing an opinion - became the touchstone for much else that followed. From discussions with the auditors[53] it became clear that the whole process would have to be documented, social accounting principles derived and stated and the accounting standards of information disclosure would have to be applied. A version of the standard long-form

8 of 18 audit report - possibly with heavy qualifications - would be appended to the social accounts. Convincing the firm of accountants to take on this responsibility was not simple and they required a great deal of "comfort" from TX. This, in turn, had a remarkably "tightening-up" effect on the whole process and certainly introduced a higher degree of independence into the process. The auditors had to be convinced by the definition of the stakeholders. This was achieved by following the same procedure described above for the plc. TX used its (very detailed) mission statement to categorise and prioritise its stakeholders to the satisfaction of the auditors[54]. Then three levels of information were collected about each relationship with each stakeholder. The stakeholder information - in the sense of an organisation seeking to manage its stakeholders - was not strictly appropriate because of TX's highly transparent culture and the existence of the detailed mission statement which acts as an effective contract between the organisation and its constituents. Information descriptive of the relationships was, however, essential and this was referred to as the "stakeholder data". The legal (identified by reference to statute law) and moral rights (as expressed in the TX mission statement) were then identified and information designed to discharge the resultant accountability was collated for each. Finally, arrangements were made, where possible through (relatively independent) third parties to seek out and collect the voices of the stakeholders. Throughout the process an audit trail was maintained whenever possible and specific notes were kept of all decisions, value judgements and missing information that (inevitably) arose during the process. The resultant social account then comprised: a section outlining the process used; the auditor's report; and the social account itself covering the six groups of stakeholders finally agreed as covering most of the material stakeholders[55]. Each stakeholder section provided the three layers of information: the descriptions of the relationship, the accountability information and the "voices" of the stakeholders. Although not all information was collected for all aspects of each of the stakeholder groups - completeness, it seems, will always be a problem - the result was a relatively transparent and replicable account. It seriously engaged the auditors and it looks highly likely that the method could be applied to any organisation. The TX social accounting process, then, has two distinctive elements: the relatively programmed approach to the collection and reporting of information; and the "true and fair" report from the statutory financial auditors. The relatively programmed approach has clearly forced the active participants in the social accounting process towards a more specific explication of the processes and judgements made. That is, although the TX experience, drawing heavily from the plc experience, was an explicitly programmed process, it was not simply mechanistic. Too many of the data and relationships were "soft" and required judgement by the social accountant. Such judgement was equally required by the auditor. The auditor's involvement forced a much more explicit process upon the accounting process and threw into relief the absence of any accounting standards or GAAP for social accounting. Whilst generally accepted auditing principles were found to be helpful in the attestation process it was an absence of anything to define the account that caused problems. (This is a subject for the final section of the paper). We could usefully express the view that, in our opinions, there never could be - nor indeed should be - a completely unqualified audit report on a social account. As a process primarily designed to enhance transparency and accountability, judgement will be essential and one should not expect an independent view to reach - or even agree with - such social judgements. As long as this is clear, the qualification of a "true and fair" report seems a desirable attribute[56]. (It is less easy to be prescriptive about the form that a PCP/Evaluator's audit report should take). In the next section we attempt to synthesis the present experience to move towards some social accounting standards.

9. Towards Social Accounting Standards It is clear that Traidcraft plc and Exchange are very unusual organisations - in their orientation, values- and transparency-based cultures and in their size. It would therefore be inappropriate to freely generalise their experience as applicable to all organisations. However, subsequent experience with different organisations (many of which are less willing to allow themselves and their inner secrets to be discussed with the same openness and freedom as Traidcraft has permitted)[57] suggests that the basic experience derived from the Traidcraft cases provides an especially useful basis on which to move towards identification of some key elements in the social accounting process. That is what this (final) section will attempt to do - but succinctly, leaving more detailed discussion of theoretical coherence of key social accounting principles to future papers. There is not, nor is there ever likely to be, a single type of social account. The range of approaches and experiments increases all the time. We consider this healthy and very desirable - not least because many different organisations in different social contexts will have different conceptions of the elements discussed here. (The range discussed in Zadek and Raynard, 1995 and the emerging NGO community-based social accounts with their own special concerns speaks eloquently of this). Furthermore, as Tinker et al (1991) have shown, no account of organisation-society interactions can be static through time - issues, concerns and accountabilities change. This is inevitable. However, an alternative approach which responds to the increasing requests for guidance on "how to do a social account" with a statement like "any way you want" is not helpful and does not necessarily advance the cause of accountability reporting in any productive way. The recent formation of the Institute for Social and Ethical Accountability (ISEA) is predicated on the need to guide organisations seeking to develop social accounting and to try and deducebest practice in the field where this is possible. This is not a simple task - nor one which can be accomplished quickly. This paper is an early attempt to begin the process of articulating some of the key matters which bodies such as the ISEA will need to address. What then, do we believe are the key elements in an approach to an "ideal type" of social account that seem to have emerged from the experiences discussed above. Figure 7 is an attempt to synthesise the messages that appear to emerge from the foregoing discussion and from the experiences of social accounting so far. In time, each of these elements will prove to be sound - and expanded upon - or will be rejected and replaced. At the present time, they appear to be key useful factors that social accounting needs to consider. Whilst Figure 7 accommodates a fairly high degree of synthesis between (i) the TX "accounting-style" approach which incorporates the stakeholder dialogue voices; and (ii) the NEF/Traidcraft approach in which the voices are paramount and the stakeholder and accountability information is incorporated; there still remains a tension about the emphasis given to these two sides of the same story. This tension is perhaps best expressed in the role attributed to the auditor or even "social Figure 7: Some Emerging Elements For Conceptual auditor". Two alternative approaches have been mentioned above. The first was the use of a Framework for CSR completely independent auditor (in the case of TX, the statutory financial auditor) who has no Conceptual pre-requisites for an involvement with the construction of the account and simple expresses a (perhaps detailed) accounting:

9 of 18 opinion on the accounts "truth and fairness". The second approach to "auditing" was the far an accounting entity; more active involvement of an organisation like NEF who help construct the social account, oversee the collection of data and responses from stakeholders and rely, in their turn, on an an accounting/account "audit advisory body" to suggest some degree of impartiality in the assessment of the Definitions derived for: account. We remain, currently, split on this issue but hope - indeed anticipate - some further synthesis of thinking on this matter as experience develops. Our current thinking suggest that social accounting; we may, indeed, have two - probably similar and complementary - constructs. The first such the recognition and construct - the one with which this paper is primarily concerned -is a formal "social account" boundaries of the accounting whose primary purpose is to construct a social image of the organisation. The second entity -via stakeholder construct is the "social audit" (which may or may not produce a social account) whose analysis; principal purpose is to encourage a negotiation for change between the organisation and its stakeholders. The precise dividing line - if indeed there is one - still needs further definition. the nature of the account and its completeness; A final matter which still needs detail specification is the way in which the stakeholder voices are "heard" and collated. These voices are not making a single statement. They are making Assumed: statements about: (a) things which the accounting organisation wishes to know about; (b) a neo-pluralist context; information which each stakeholder would like to receive; and (c) their views (for example, complaints) on the organisation and its activities plus the activities and issues the that a stakeholder relationship stakeholders would like the organisation to address. The first of these seems to relate to the conception of an organisation stakeholder perspective and be useful for organisational management. The second seems to reflects a neo-pluralist relate to accountability and the stakeholders determining the terms of accountability that they assumption. would wish. In both of these cases, it seems likely that the data could be collected by, inter desirability of democracy; alia, questionnaires. Confidentiality does not appear to be so crucial here. The third type of accountability as a democratic information does, however, require to be collected and collated - and reported? - by an mechanism; entirely independent unit which the stakeholder can trust. (Otherwise are, for example, employees going to say things which might incur the wrath of management?). This is an organisational self-reporting emerging issue on which more work needs to be done. (as opposed to "social audits"); The social accounting agenda has opened out as never before. Certainly, for the first time in more than 20 years there is an active dynamic process with which researchers and teachers information should possess need to engage. Academic engagement with the environmental accounting and reporting the characteristics of process has been fairly successful. Drawing from that, academe now needs to help bodies completeness, reliability, like ISEA set a new foundation for the formal accountability of organisations. It is an verifiability, consistency, important, exciting and very urgent challenge. comparability, understandability; Appendix 1: Traidcraft plc's experience with one of its stakeholder groups the account social The early stages of the Traidcraft plc experience led us to believe that each of the identified re-constructs an organisation relationships needed to be described in some way which reflected the understanding of the which pre-exists the account.; parties to that relationship - noting that the way in which they are accounted for will involve The re-construction should be financial, quantitative and descriptive elements. It is also likely that the way in which the a reflexive and complex series descriptors of the relationships are derived will be partly pragmatic - based upon what can be of views from all stakeholders. described and what data can be collated. Deduced: For illustration, we run through the a priori analysis of what (we believe) is the most important of the Traidcraft relationships - the one they hold with their developing country producers. groups and individuals have rights to information; The Producers (i.e. Third World Suppliers) these are resolved informally This is the dominant social relationship of Traidcraft plc. It is Traidcraft's raison d'etre without in smaller non-complex which Traidcraft would not exist as it is currently conceived. This must, therefore, be one of situations; the dominant areas of the social account. The major elements of the relationship are derived formal accounts are needed from Traidcraft's Purchasing Policy and the resultant elements are appropriately listed in the for accountability in complex 1993 Social Accounts. In addition, any account of the relationship with the producers needs situations; (i) to be set in a context (of the range of suppliers, the products and the existing living/working conditions pre and post-Traidcraft) and (ii) to explore the implied terms of the relationship - in organisations need to produce particular the need to be in harmony with the local culture and to develop a closeness of accountability accounts; relationship. rights to information For each characteristic, the trick is to try and capture the essence of that characteristic and determined by (a) the then to allow the producers themselves to speak about the characteristics of the relationship. organisation (b) society (c) the The second part of this was done through direct (face-to-face) consultation with a selection of stakeholders; [58] the producers and backed up by a distant questionnaire. The results of this are shown in the three layers of information the Traidcraft plc Social Accounting Statements. The first of the two elements above - -plus the descriptors of the capturing the principle descriptors of the relationships - formed the basis of the social relationships must be a bookkeeping system (see Dey et al. 1995). Our first attempt to articulate these descriptors - "complete" account or along with sources of information for the social account (that is, the means of social information provided to assess bookkeeping) - are shown in Figure A1. the extent of the FIGURE A1 ABOUT HERE incompleteness; independent judgement - References and Bibliography preferably inexperienced in social accounting - is essential Accounting Standards (formerly Steering) Committee (1975) The Corporate Report (London: to assess completeness and ICAEW) reliability Adams R. (1989) Who Profits? (Oxford: Lion) Asserted: American Institute of Certified Public Accountants (1977) The measurement of corporate

10 of 18 social performance (New York: AICPA) social accounting is a dynamic Arrington E. (1990) "Intellectual Tyranny and the Public Interest: The quest for the holy grail and developing process. and the quality of life" Advances in Public Interest Accounting Vol.3 (pp1-16) Nothing here should limit further development and Baxter W.T. (1988) Academic Research: Academic trends versus professional needs experimentation; (Edinburgh: ICAS) the social accounting is a Beresford D.R. and L.T.Johnson (1995) "Interactions between the FASB and the academic continuous process of iteration community" Accounting Horizons 9(4) December (pp108-117) and negotiation. Cooper C. (1992) "The Non and Nom of Accounting for (M)other Nature" Accounting, Auditing and Accountability Journal 5(3) (pp16-39) Dey C., R.Evans & R.H.Gray (1995) "Towards social information systems and bookkeeping: A note on developing the mechanisms for social accounting and audit" Journal of Applied Accounting Research 2(3) December Donaldson J. (1988) Key Issues in Business Ethics (London: Academic Press) Dowling J. & J. Pfeffer (1975) "Organisational Legitimacy: Social values and organisational behaviour " Pacific Sociological Review January (pp122-136) Epstein E.M. & D. Votaw (1978) (ed) Rationality, Legitimacy, Responsibility: Search for new directions in business and society (California: Goodyear) Estes R.W. (1976) Corporate Social Accounting (New York: Wiley) Evans R. (1991) "Business Ethics and Changes in Society" Journal of Business Ethics 10 (pp871-976) Francis J. (1990) "After virtue? Accounting as a moral and discursive practice " Accounting, Auditing and Accountability Journal 3(3) (pp5-17) Gallhofer S. and J. Haslam (1995) "Worrying about environmental auditing" Accounting Forum 19(2/3) September/December (pp205-218) Gallhofer S. and J. Haslam (1993) "Approaching Corporate Accountability: Fragments from the past" Accounting and Business Research 23(91a) pp320-330 Geddes M. (1988) Social Audits and Social Accounting: An annotated bibliography and commentary School of Applied Economics and Social Studies, South Bank Polytechnic, London. Geddes M. (1991) "The social audit movement " in D.L.Owen (ed) Green Reporting (London: Chapman Hall) (pp215-241) Gray R.H. (1989) "Accounting and democracy " Accounting, Auditing and Accountability Journal 2(3) (pp52-56)

Gray R.H. (1992) "Accounting and environmentalism: an exploration of the challenge of gently accounting for accountability, transparency and sustainability " Accounting Organisations and Society 17(5) July (pp399-426). Gray R.H. (1996) "The interesting relationship between accounting research and accounting practice: A personal reply to Professor Whittington" Journal of Applied Accounting Research March 1996 Gray R.H. (forthcoming) "The silent practice of social accounting and corporate social reporting in companies" in Zadek S., R.Evans & P.Pruzan (eds) Challenging the Corporate Soul: The practice of social and ethical accounting (London: Earthscan). Gray R.H., K.J.Bebbington, D.Walters & I.Thomson (1995b) "The Greening of Enterprise: An exploration of the (non) role of environmental accounting and environmental accountants in organisational change " Critical Perspectives on Accounting 6(3) (pp211-239) Gray R.H., R.Kouhy & S.Lavers (1995a) "Corporate social and environmental reporting: A review of the literature and a longitudinal study of UK disclosure " Accounting, Auditing and Accountability Journal 8(2) (pp47-77) Gray R.H., R.C.Laughlin & K.J.Bebbington (1996b) Financial Accounting: method and meaning (London: International Thomson) Gray R.H., D.L. Owen & C.Adams (1996a) Accounting and accountability: Changes and challenges in corporate social and environmental reporting (London: Prentice Hall) Gray R.H., D.L.Owen & K.T.Maunders (1986) "Corporate social reporting: the way forward?" Accountancy December (pp6-8) Gray R.H., D.L.Owen & K.T.Maunders (1987) Corporate Social Reporting: Accounting and accountability (Hemel Hempstead: Prentice Hall) Gray R.H., D.L.Owen & K.T.Maunders (1988) "Corporate social reporting: emerging trends in accountability and the social contract " Accounting, Auditing and Accountability Journal 1(1) (pp6-20) Gray R.H., D.L.Owen & K.T.Maunders (1991) "Accountability, Corporate Social Reporting and the External Social Audits " Advances in Public Interest Accounting Vol.4 (pp1-21) Grinyer J. (1996) "Perceptions of financial managers concerning their role and the relevance of research " in Tomkins C., Grinyer J., McAulay L., Osbourne P., & Walker M. (eds) In Search of Relevance for Financial Management Research British Accounting Association Monograph, (pp19-33) Guba E.S. & Y.S.Lincoln (1989) Fourth Generation Evaluation (Newbury Park, Calif: Sage) Habermas J. (1978) Knowledge and Human Interests (Londonm: Heinemann) Harte G. & D.L. Owen (1987) "Fighting de-industrialisation: the role of local government social audits " Accounting, Organizations and Society 12(2) (pp123-142)

11 of 18 Held D. (1987) Models of Democracy (Oxford: Polity Press) Hines R.D. (1988) "Financial accounting: In communicating reality, we construct reality " Accounting, Organizations and Society 13(3) (pp251-261) Hines R.D. (1989) "The sociopolitical paradigm in financial accounting research " Accounting, Auditing and Accountability Journal 2(1) (pp52-76) Hines R.D. (1991) "The FASB's Conceptual Framework, Financial Accounting and the Maintenance of the Social World " Accounting, Organizations & Society 16.4 (pp313-332) Hines R.D. (1991) "Accounting for Nature" Accounting, Auditing and Accountability Journal 4(3) (pp27-29) Hines R.D. (1992) "Accounting: Filling the Negative Space " Accounting, Organizations and Society 17(3/4) (pp313-342) Hopwood A. Accounting Research and Accounting Practice: The ambiguous relationship between the two Deloitte Haskins & Sells Accounting Lectures, UCW Aberystwyth, 1983 Hutchins B. (1987) "A day's work" AA(Accountancy Age) August p48 Innes J., W.A.Nixon and N.Tagoe (1996) "Accounting and Strategic Processes - A Case Study" University of Dundee Discussion Paper ACC/9601 Laughlin R.C. (1991) "Environmental Disturbances and Organisational Transitions and Transformations: Some Alternative Models " Organisation Studies 12(2) (pp209-232) Laughlin R.C. & J.Broadbent (1996) "Redesigning Fourth Generation Evaluation: An evaluation model for the public sector reforms in the UK?" Evaluation 2(4) (pp431-451) Lavoie D. (1987) "The accounting of interpretations and the integration of accounts: The communication function of the language of business" Accounting, Organizations and Society 12(6) (pp579-604) Lazarsfeld P.F. (1971) "Accounting and social bookkeeping" in R.R.Sterling & W.F.Bentz (eds) Accounting in Perspective (Houston: Scholars Books) pp88-101 Lee T.A. (1989) "Education, practice and research in accounting: Gaps, closed loops, bridges and magic accounting " Accounting and Business Research Summer (pp237-254) Lehman G. (1995) "A legitimate concern for environmental accounting " Critical Perspectives on Accounting 6(6) (pp393-412) Likierman A. (1986) Rights and obligations in public information (Cardiff: University College Cardiff Press) Likierman A. & P. Creasey (1985) "Objectives and entitlements to rights in government financial information" Financial Accountability and Management 1(1) Summer (pp33-50) Llewellyn S. (1994) "Managing the boundary: How accounting is implicated in maintaining the organisation" Accounting Auditing and Accountability Journal 7(4) (pp4-23) Macpherson C.B. (1973) Democratic Theory: Essays in retrieval (Oxford: OUP) Macpherson C.B. (1977) The life and times of liberal democracy (Oxford: OUP) Malachowski A. (1990) "Business ethics 1980-2000: an interim forecast " Managerial Auditing Journal 5(2) (pp22-27) Mathews M.R. (1993) Socially Responsible Accounting (London: Chapman Hall) Maunders K.T & R. Burritt (1991) "Accounting and Ecological Crisis" Accounting, Auditing and Accountability Journal Vol 4, No 3, pp. 9-26. Medawar C. (1976) "The social audit: a political view" Accounting, Organizations and Society 1(4) (pp389-394) Meyer J.W. (1986) "Social environments and organisational accounting " Accounting, Organizations and Society 11(4/5) (pp345-356) Mitchell A., A.Puxty, P.Sikka & H.Willmott (1991) Accounting for change: Proposals for reform of audit and accounting (London: Fabian Society) Nasi J. (ed) (1995) Understanding Stakeholder Thinking (Helsinki: LSR Publications) Owen D.L. (1997, forthcoming) "Social and Environmental Accounting and Sustainability: Keynote Speech to the EMAA Conference" Proceedings of the EMAA Conference (Amsterdam: Limperg Instituut) forthcoming Pallot J. (1987) "Infrastructure assets as a concept in governmental accounting " paper given to AAANZ Canberra, Pallot J. (1991) "The legitimate concern with fairness: A comment " Accounting Organisations and Society 16(2) (pp201-208) Parker L.D. (1986) "Polemical themes in social accounting: A scenario for standard setting " Advances in Public Interest Accounting Vol.1 (pp67-93) Pearce J. (1993) At the heart of the community: Community Enteprise in a changing society (London: Caloustie Gulbenkian Foundation) Pearce J. (1996) Measuring Social Wealth: A study of social audit practice for community and cooperative enterprises (London: New Economics Foundation) Power M. (1994) "Constructing the responsible organisation: Accounting and environmental representation" Chapter 16 in G.Teubner, L. Farmer and D.Murphy (eds) Environmental Law and Ecological Responsibility: The concept and practice of ecological self-organisation (London: John Wiley) pp370-392 Puxty A.G. (1986) "Social accounting as immanent legitimation: A critique of a technist ideology " Advances in Public Interest Accounting Vol.1 (pp95-112) Puxty A.G. (1991) "Social accountability and universal pragmatics " Advances in Public Interest Accounting Vol.4 (pp35-46)

12 of 18 Raynard P. (1995) "The New Economics Foundation's Social Audit: Auditing the auditors?" Social and Environmental Accounting 15(1) 1995 (pp7-10) Roberts J. (1991) "The possibilities of accountability " Accounting, Organizations and Society 16(4) (pp355-370) Roberts J. & R.Scapens (1985) "Accounting Systems and Systems of Accountability " Accounting, Organizations and Society 10.4 (pp443-456) Roberts R.W. (1992) "Determinants of corporate social responsibility disclosure " Accounting, Organizations and Society 17(6) (pp595-612) Sikka P. (1987) ' Professional Education and Auditing Books: A Review Article' British Accounting Review Vol 19, No 3, December pp. 291-304. Sikka P., H.C. Willmott & E.A. Lowe (1989) ' Guardians of Knowledge and the Public Interest: Evidence and Issues of Accountability in the UK Accountancy Profession' Accounting, Auditing and Accountability Journal Vol 2, No 2, pp. 47-71. Sikka P., H.C.Willmott & E.A.Lowe (1991) "'Guardians of knowledge and the public interest': A Reply to our critics " Accounting, Auditing and Accountability Journal Vol 4, No 4, pp. 14-22. Sikka P., H.Willmott and T.Puxty (1995) "The mountains are still there: accounting academics and the bearings of intellectuals " Accounting, Auditing and Accountability Journal 8(3) (pp113-140). Sterling R.R. (1973) "Accounting research, education and practice " Journal of Accountancy September (pp44-52) Stone C.D. (1975) Where the law ends (New York: Harper & Row) Tinker A.M., Lehman C. & Neimark M. (1991) "Corporate Social Reporting: Falling down the hole in the middle of the road " Accounting, Auditing & Accountability Journal 4(1) (pp28-54). Tomkins C., Grinyer J., McAulay L., Osbourne P., & Walker M. (eds) In Search of Relevance for Financial Management Research British Accounting Association Monograph, forthcoming Tricker R.I. (1983) "Corporate responsibility, institutional governance and the roles of accounting standards " In Bromwich M. & A.G.Hopwood (eds) Accounting Standards Setting - An International Perspective (London: Pitman) Ullmann A.E. (1985) "Data in Search of a Theory: A Critical Examination of the Relationships Among Social Performance, Social Disclosure and Economic Performance of US Firms " Academy of Management Review 10(3) (pp540-557) Whittington G., (1995) Is Accounting Becoming too Interesting? Sir Julian Hodge Lecture, University of Aberystwyth, pp. 20. Williams P.F. (1987) "The legitimate concern with fairness " Accounting Organisations and Society 12(2) (pp169-189) Zadek S. (1993) "The Social Audit of Traidcraft plc " Social and Environmental Accounting 13(2) (pp5-6) Zadek S. & R. Evans (1993) Auditing the Market:A Practical Approach to Social Accounting (London: New Economics Foundation) Zadek, S., P.Pruzan & R.Evans (forthcoming) Enhancing corporate social performance: The practice of social and ethical accounting, auditing and reporting (London: Earthscan) Zadek S. and P.Raynard (1995) "Social auditing, transparency and accountability" Accounting Forum 19(2/3) September/December

13 of 18

[1] We gratefully acknowledge the assistance provided by many individuals and organisations in developing this paper. We are grateful to the financial support received from the ESRC, the Carnegie Trust and the ACCA who have funded - or are continuing to fund - elements of this research project. We wish to express our thanks to the staff at Traidcraft plc and Traidcraft Exchange (especially Murdoch Gatwood) who have been so supportive and patient throughout the (continuing) research process. Jan Bebbington, David Collison, Emily Dick-Forde, Reza Kouhy, Ken McPhail and Noel Tagoe of The University of Dundee and Sue Llewellyn of Edinburgh University were especially helpful in helping us try and clarify the theoretical issues raised by the paper. We also acknowledge the very helpful comments on earlier drafts of this paper from Nabil Baydoun, Jan Bebbington, Jane Broadbent, Penny Ciancanelli, John Grinyer, Jim Haslam, John Innes, Richard Laughlin, Alasdair Lonie, Peter Raynard and colleagues who attended presentations at the CSEAR Summer School, The University of Glasgow and the University of Dundee.

[2] The issue of sustainability makes this a contestable suggestion. That is, if social (and environmental) accounting are re-conceived within the concept of sustainability then broad (deep-green) conceptions of the environment will provide an over-arching framework for the conceptions of social accounting (to deal with justice issues) and environmental accounting (to deal with resource and physical environmental issues). The project suggested by such a re-conception is still in a very early stage of development. It perhaps also worth emphasising that if sustainability is to become the central policy principle guiding future economic development and accounting can play any role in this, social accounting may well be able to act for eco-justice the way environmental accounting is acting for eco-efficiency. See, for example, Owen (forthcoming).

[3] Convincing histories which explain the rise of environmental accounting and, more recently, of social accounting, have yet to be

14 of 18 written.

[4] British Telecom, one of Britain's largest companies, has publicly stated its commitment to produce and publish a social account. Other large companies are actively exploring these possibilities themselves - but in a less public manner.

[5] Contact any of the authors for more information about the Institute.

[6] There is a frequently voiced concern that academic accounting fails to engage sufficiently with the practice of accounting. This concern, which we accept as being broadly legitimate, is not, however, the product of a single voice. Practitioners express concerns that they neither understand - nor can see the applicability of - much accounting research (see, for example, Grinyer, 1996; Tomkins et al, 1996). Accounting policy-makers - themselves often academics - bemoan the lack of useful guidance offered by the research community, (see, for example, Sterling, 1973; Baxter, 1988; Lee, 1989; Beresford and Johnson, 1995; Whittington, 1995). Radical researchers accuse academics of failing to challenge practice and to exercise their political and ethical duty to seek to change practice, (see, for example, Sikka, 1987; Sikka et al., 1995). Indeed, many radical researchers, in their criticisms of practice, could be thought to be suggesting that much practice is fundamentally flawed and, consequently, that seeking to engage with an agenda already controlled by practice - and corporate capitalism - is largely a waste of time. This last criticism, whilst well-founded, offers little scope for praxis.

[7] The relationship between research and practice is clearly not an ideal one. The relationship is, however, a complex one, (see, for example, Hopwood, 1983; Gray, 1996) in which different sections of the practice and academic communities hold significantly different objectives for the practice of accounting and make significantly different judgements about the responsibility of practitioner and academic to seek change. But part of this complexity arises from matters of practicability and pragmatism. As a backdrop to this paper, imagine that all accounting practice was abolished overnight and the academic community were asked to start with a clean sheet and redesign a full accounting practice from scratch. There is virtually nothing in the literature which systematically addresses how one might construct a "utopian" accounting. The suspicion must be that (a) accounting academics themselves would find such a challenge "unrealistic"; (b) that such analysis might fail to satisfy the current output measurement system and hierarchical pecking order which govern academe these days; and (c) that any such attempt would be heavily influenced by current practice anyway. So, in a very general sense, we can see academe as an implicitly negotiated balance between current practice, pragmatic opportunities, and "more desirable" - but practicable areas of change.

[8] This is the sense in which praxis is used here.

[9] As we will discuss below, the democracy referred to here is participatory democracy in a neo-pluralist conception, (see, for example, Held, 1987; Macpherson, 1973;1977; Gray, 1989). We should stress that we are aware of the potential cultural specificity of this basis and, therefore, our discussion relates, principally, to European and Anglo-American countries plus those nations strongly influenced by those cultures. We will stress below that we see accounting and social accounting as culturally specific and contextually determined.

[10] All the authors have been involved, to varying degrees, in the production of the UK social accounts, in the RSA enquiry and in the development of the ISEA.

[11] There are other important developments in social accounting which are not addressed here. The most significant relates, in the UK at least (but see also Zadek and Raynard, 1995), to producing social accounts for "community enterprises" or "third sector". Here, the exigencies and motivations are somewhat different from the commercial sector. Nevertheless, the social accounting models are developing with a useful degree of overlap with the commercial sector. See, for example, Pearce (1993) and John Pearce's new book on social accounting practice and social bookkeeping systems which is to be published by the New Economics Foundation in due course. The exclusion of these works from our discussion is simply expedient and cautious. We see very strong sympathies and synergies between these two areas of social accounting development and ultimately would like to see social accounting as universal for all organisations.

[12] We are explicitly conscious of the ambition in this. Two thousand years of conventional accounting have failed to produce a (universally accepted) conceptual framework. It is somewhat unlikely that a single paper (or even series of papers) will succeed in this endeavour for the more complex case of social accounting. However, it is appropriate to note that (i) social accounting, unlike conventional accounting, does not yet have a wide legislative basis from which to build and (ii) the derivation of a social accounting from the foundations of conventional accounting thinking may be mis-guided. These issues have not been explored in this paper.

[13] For further detail see, for example, Held (1987); Macpherson (1973;1977); Gray (1989); Gray et al. (1996).

[14] These views and assumptions are informed, in the very broadest terms by the works of, inter alia, Habermas, Rorty, Rawls and Dewey. Within the accounting literature, we would refer the reader to the work of Arrington, (see, for example, Arrington, 1990), Broadbent and Laughlin, (see, for example, Laughlin and Broadbent, 1996) and Lehman, (see, for example, Lehman, 1995). For a different conception of related issues, see, for example, Gallhofer and Haslam, (1993).

[15] See Zadek and Raynard (1995) who present a range of recent examples of social accounting which, whilst broadly similar in orientation, employ quite different terminology.

[16] For example, personal experience suggests that the radical socialist overtones of "social accounting" in the 1970s (see, for example, Gray et. al. 1996), the radical activism of the "social audits" of the 1980s (see, for example, Geddes, 1991; Harte and Owen, 1987) and the widespread British distaste for matters social (e.g. the Social Chapter) have led to a significant distrust of anything with "social" in the title amongst major pockets of British political and industrial power. Similarly, the language of accounting does not always appeal to social accounting practitioners. Experience suggests that the analogue with, especially, independent audit (see below) has negative connotations for many business people due to their less than overwhelming enthusiasm for the accounting profession and their negative experiences with the (far from) independent audit process.

[17] Implicit in this reasoning is a recognition than whilst there is a considerable amount of conventional accounting that is problematic, many of its central functions and tenets are well-established, highly theorised and practicable. (If, that is, one has accepted the need for any form of formal accounting at all). We are seeking here to employ the "best" from conventional accounting and throw away the rest - which would comprise, almost certainly, much of the detail of conventional financial accounting.

[18] Organisational is used here as a generic term. Whilst we wish to restrict our engagement here to, predominantly, large organisations,

15 of 18 we do not necessarily wish to exclude, for example, families and individuals. The term "organisation", in a systems context, can be taken to embrace all "units". (For more detail, see Gray et al., 1996b)

[19] Although, as we shall see, the relationship is a reflexive one.

[20] This point, as we saw earlier, is taken as axiomatic in the present conception of accountability.

[21] This paper will not be examining the pros and cons of social accounting. This has been extensively covered elsewhere. See Gray et al. (1996a) for a summary.

[22] The use of the term "pragmatism" in this paper is the normal English Language usage of the term and is related in only the most tenuous way to the work of, for example, Dewey and his colleagues.

[23] And, indeed, to the development of independent reporting under the European EMAS environmental reporting scheme.

[24] Not because we consider business' views on desirability are a dominant criterion but rather to highlight where the conflict from social accounting is more likely to arise.

[25] It should be apparent that the issue of why an organisation is undertaking a social accounting is not discussed here. In addition to the problems that inferring motivation brings, we are generally working in a normative vein in which it is assumed that the motive should be the discharge of accountability. It is the nature of that accountability that is the issue.

[26] We should emphasise that additional theories can be found and, in particular, those theories informed by marxian thinking would largely challenge our approach here. We do not wish to dispute the marxian approach here but in attempting to develop a systematic mechanism for social accounting by the entity itself, a conception which challenged the entity's very right of existence is unlikely to seem fruitful to the organisation concerned. We are, therefore, choosing to work in a neo-pluralist conception of the organisation-society interface (see, for example, Held, 1987; Gray et.al., 1991; 1996) with all the attendant problems that that attracts from a critical perspective (see, for example, Tinker et al, 1991 but see also Lehman, 1995). For the sake of completeness we should mention that the broad accounting literature does contain very many other theoretical perspectives within which an organisation-society relationship is often implied. The "economic-based" theories (see, for example, Gray, Kouhy and Lavers, 1995a) of agency theory, decision-usefulness and information economics theory assume a narrower conception of both the relationship, the motives of the parties and the "society" involved. They respond more economic power than to rights and wider notions of justice and accountability, (see, for example, Lehman, 1995). Other theories, for example, legitimacy theory and political economy theory seem to be more appropriate for the analysis of existing practices than as normative bases from which to deduce "proper" accountability relationships, (see, again, Gray, Kouhy and Lavers, 1995a).

[27] As far as we are aware, this approach has yet to enter management thinking - at least in the form in which it is employed here - in any systematic way. Hence the (somewhat clumsy) coined phrase.

[28] This focus on groups will be continued throughout the paper and helps explain our attachment to neo-pluralist thinking.

[29] We would like to acknowledge the work of David Woodward in this field. His papers and the conversations we have held have been most helpful in guiding us towards a clearer idea of the elements of active stakeholder thinking. See also Nasi (1995).

[30] This point is also made by Tricker (1983) in his arguments that real accountability is positive accountability - that is, if an accountee cannot enforce an accountability disclosure then there is no real accountability relationship.

[31] Indeed, an important part of the social accounting approach taken by Traidcraft and the New Economics Foundation - of which more below - is the establishment of the organisation's core values, how the organisation performs against those core values and the views of the stakeholders on those values and the performance.

[32] "Good" information is typically assumed to have a number of qualitative characteristics such as reliability, completeness and so on, (see, for example, ASSC, 1975). To satisfy these characteristics (which are returned to in the closing section of the paper) it seems practically necessary to also provide data which can be thought of as describing the terms of the relationship between the organisation and its stakeholders. This descriptive data on, for example, numbers and types of employees, sources and types of suppliers, nature of the organisation's interactions with the natural environment, and so on, seems to be a necessary element within the information disclosure to provide context for the accountability information so reported.

[33] See, Laughlin and Broadbent (1996) for an especially clear critique of Guba and Lincoln in which that work is contrasted with the insights of Habermas and a Habermasian development and synthesis is attempted.

[34] Whilst this alternative conception may come in various forms (see, for example, Zadek and Raynard, 1995) and has significant overlap with aspects of the (characterisation of) the stakeholder and accountability perspectives offered above, (as we shall see), its differences are important.

[35] The problem of how to give "silent" stakeholders - typically the unborn/future generations, the environment and, if one is to avoid anthropocentrism, non-human life - a voice is still unresolved.

[36] We are grateful to Ken McPhail for his help on this issue. Ken argues that the value of the postmodern project lies, not in the existence of a "postmodern" epistemology so much as, in the tension and critique of modernism offered by that project. Therefore we might see that, to varying degrees, the Habermasian, Foucauldian and Guba and Lincoln projects work at the point of this tension.

[37] What we see, therefore, is a more fluid construction of an accounting entity which explicitly recognises parties other than the managers of the organisation in the construction of the organisation. In this it is somewhat akin to Llewellyn's construction of the "postmodern" accounting entity, (Llewellyn, 1994). Secondly, in the privileging of the stakeholders' voices, we have a reversal of the traditional stakeholder theory which, as we have seen, privileges the organisation's voice. Finally, we see that the process is hermeneutic, dynamic and iterative.

[38] The Guba and Lincoln requirement that evaluation leads to action is also central to the PCP approach outlined here.

16 of 18 [39] The limits of an empirical accountability lie, inter alia, in the extent to whichpositive rights - i.e. those which are demonstrable - successfully reflect a wider democratic awareness or, more likely, reflect un-democratic power structures. (See, for example, Stone, 1975; for one excellent discussion of this and related issues). This role - of extending beyond the empirical accountability - was previously ascribed to the "external social audits" (see Figure 1; Gray et al, 1988; Harte and Owen, 1987; Geddes, 1988) - a role which we can see social audits continuing to play as a complement to the present social accounts.

[40] This echoes a common concern that arises when constructing social accounts with the cooperation of the organisation. That is, any organisation, it seems, will want information for control purposes whilst the principal objective of the social account is accountability purposes. Whilst the information will usually be (and may frequently be perceived as being - see, for example, Roberts and Scapens, 1985) seen as comprising two overlapping sets, there will be many occasions when the demands of control information will conflict with the demands for accountability information.

[41] Like conventional social accounting, there is a danger that the process will be captured by capital. But, unlike the more conventional approaches, the reader of such an account would have no external yardstick against which to assess the degree of that capture. This may be a role for the "social auditor" - a matter to which we return later.

[42] One essential issue which has yet to be addressed by the PCP approach is that of "completeness" of the resultant report. This and related matters are dealt with later in the paper but we can see at this stage that should an organisation find itself recording many seriously dissatisfied voices the organisation may very reasonably be disinclined to continue with the accounting on the grounds that (a) it is not an essential part of accountability and (b) it is difficult for an organisation to justify expenditure in reporting all views expressed by its critics. Whilst, in a perfectly transparent world we may well desire such a situation, pragmatism and a recognition of current organisational constraints suggest that such ambitions lie beyond practicality at the moment.

[43] Copies of the Traidcraft social audits can be consulted in CSEAR's library and are available from Richard Evans, Traidcraft Exchange, Kingsway North, Gateshead, Tyne and Wear, NE11 0NE, UK.

[44] This was published following a document in 1992 "Towards a Social Audit" in which the early ideas and principles were discussed.

[45] See also Hutchins (1987) and Adams (1989) for more historical expressions of the company's position.

[46] The 1993 report was sent to all the UK's top 100 companies asking for comments and enquiring whether the company might not like to undertake a similar exercise.

[47] There was confusion, for example, over whose point of view was to be taken (e.g. the company's or society's) and how this was to be established. There was also confusion over how the parameters of the conception were to be determined. In this, Traidcraft's experience echoes that of the early social accounting experiments, (see, for example, Gray et al., 1987). The discussion earlier in this paper is both motivated and informed by the process of resolving this confusion.

[48] Other issues had also arisen. In particular, the advisory board had recognised the need for a systematic system of "social bookkeeping" - a term it had coined to relate to the data which fed into a specified social account. (But see Lazarsfeld, 1971; for an earlier use of the term). A discussion and exploration of this is presented in Dey, Evans and Gray (1995). Furthermore, a growing literature was examining, not just how accountability works and can be conceived of (see, for example, Gray et al., 1987; 1988; 1991; Roberts, 1991; Roberts and Scapens, 1985; Pallot, 1991; Williams, 1987; Tinker et al., 1991), but what changing forms of accountability meant to the organisation and its participants, (see, for example, Llewellyn, 1994; Laughlin, 1991; Gray et al, 1995b). These latter issues are the subject of continuing research by one of the present authors.

[49] Therefore, any attempt at social accounting must, inevitably, be derived from first principles and will, almost certainly, have to evolve iteratively as the weaknesses, omissions and biases in the account are refined. Because of this, the basis on which the social accounting is developed would, ideally, be as transparent as possible with the (inevitable) political, sociological and construction biases exposed as clearly as possible.

[50] The simplification arises for a number of reasons. A simple Venn diagram would have to work in a number of dimensions to capture the overlap and inter-dependencies in the relationships. Also, the diagram implies a series of static relationships when, in practice, these are continually fluid.

[51] This is justified politically by recognition that all active parties in the organisation effectively 'contract' with the organisation and the basic terms of that social contract are the explicit elements of the mission statement.

[52] Of some importance are the two dimensions: (a) the formal/informal dichotomy in each relationship and (b) the extent to which an account is an account to the party and the extent to which the account is an account of the relationship. Each of these will need further thought than we can give them here but, it seems currently plausible that (a) the existence of the social account is a formal account - the organisation and the parties will be informally accounting to each other continually and this will not be formalised. It will however, (relating to dichotomy (b)), be necessary to account for both the informal and informal aspects of the relationship. And, it seems at this stage, this is best done through an account of the relationships. Whether correct or not, the following proceeds on the basis of formal accounts of the relationships.

[53] Rainbow Gillespie and Co, Gateshead are a small local firm of accountants which has long audited the financial statements of the Traidcraft organisations but has had no previous involvement with social accounting.

[54] The TX process here drew heavily from the plc experience described above. This was also the first TX social account and, also innovative in a number of ways. It seemed apposite therefore to simplify the definition of the accounting entity and let this iterate over subsequent social accounting cycles.

[55] There is still further discussion to be held about completeness and materiality.

[56] The issue of whether or not the "true and fair" audit report is the most suitable for the attestation of the social account needs further exploration. This form was used here in order to explore the extent to which conventional accounting experience could be used as a template for the social accounting and auditing process.

17 of 18 [57] Simon Zadek, at the time of writing is spending an increasing amount of his time simply discussing and undertaking various roles within social accounting processes and Richard Evans has been made Director for Social Accounting at Traidcraft with intention of acting as a consultancy to organisations seeking to develop their own social accounts. They both - together with Rob Gray and others - are instrumental in the formation and development of the ISEA

[58] There is a significant pragmatic, economic limit on this consultation. Travel overseas involves significant funds and time.

© CSEAR, School of Management - University of St Andrews - The Gateway, North Haugh, St Andrews, KY16 9RJ. Scotland, UK Tel: +44 (0)1334 46 2805; Email: [email protected] The University of St Andrews is a charity registered in Scotland, No SC013532.

18 of 18